I've searched the internet far and wide (for at least half a day now) and I can't seem to find the answers needed.
Currently I'm trying to create a .bnf-file for an IntelliJ-Plugin with custom language support.
A few tutorials mention the existance of {pin=1},{pin=2} and {recoverWhile=xyz}, but I didn't find any real explanation on their uses, and if there are any other things I should know (maybe a {pin=3} also exists?).
So could somebody tell me what exactly those flags, methods or however they're called are, and how to use them in my .bnf, please?
Thank you for your help and best regards,
Fuchs
These attributes are explained here:
https://github.com/JetBrains/Grammar-Kit/blob/master/HOWTO.md#22-using-recoverwhile-attribute
https://github.com/JetBrains/Grammar-Kit/blob/master/TUTORIAL.md
But the usage is not trivial. A good idea is to use Live Preview to play around with it.
My understanding:
Pin and recoverWhile attributes are used to recover parser from errors.
Pin specifies a part of the rule (by index or literally) after successful parsing of which the rule considered successful.
In the example:
expr ::= expr1 "+" expr2 {pin=1}
if expr1 is matched, the whole rule will be considered successful and parser will try yo match the rest.
if pin=2 the rule will be considered successful after matching "+" and will fail if expr1 or "+" not matched.
RecoverWhile attribute specifies where to skip after parsing the rule. Independently of its success.
For example
{recoverWhile=expr_recover}
expr_recover ::= !(";" | ".")
will skip all input before ";" or ".". I.e. parser will start matching next rule from ";" or ".".
Related
Let's say I'm writing a parser that parses the following syntax:
foo.bar().baz = 5;
The grammar rules look something like this:
program: one or more statement
statement: expression followed by ";"
expression: one of:
- identifier (\w+)
- number (\d+)
- func call: expression "(" ")"
- dot operator: expression "." identifier
Two expressions have a problem, the func call and the dot operator. This is because the expressions are recursive and look for another expression at the start, causing a stack overflow. I will focus on the dot operator for this quesition.
We face a similar problem with the plus operator. However, rather than using an expression you would do something like this to solve it (look for a "term" instead):
add operation: term "+" term
term: one of:
- number (\d+)
- "(" expression ")"
The term then includes everything except the add operation itself. To ensure that multiple plus operators can be chained together without using parenthesis, one would rather do:
add operation: term, one or more of ("+" followed by term)
I was thinking a similar solution could for for the dot operator or for function calls.
However, the dot operator works a little differently. We always evaluate from left-to-right and need to allow full expressions so that you can do function calls etc. in-between. With parenthesis, an example might be:
(foo.bar()).baz = 5;
Unfortunately, I do not want to require parenthesis. This would end up being the case if following the method used for the plus operator.
How could I go about implementing this?
Currently my parser never peeks ahead, but even if I do look ahead, it still seems tricky to accomplish.
The easy solution would be to use a bottom-up parser which doesn't drop into a bottomless pit on left recursion, but I suppose you have already rejected that solution.
I don't understand your objection to using a looping construct, though. Postfix modifiers like field lookup and function call are not really different from binary operators like addition (except, of course, for the fact that they will not need to claim an eventual right operand). Plus and minus intermingle freely, which you can parse with a repetition like:
additive: term ( '+' term | '-' term )*
Similarly, postfix modifiers can be easily parsed with something like:
postfixed: atom ( '.' ID | '(' opt-expr-list `)` )*
I'm using a form of extended BNF: parentheses group; | separates alternatives and binds less stringly than concatenation; and * means "zero or more repetitions" of the atom on its left.
Another postfix operator which falls into the same category is array/map subscripting ('[' expr ']'), although you might also have other postfix operators.
Note that like the additive syntax above, selecting the appropriate alternative does not require looking beyond the next token. It's hard to parse without being able to peek one token into the future. Fortunately, that's very little overhead.
One way could be for the dot operator to parse a non-dot expression, that is, a rule that is the same as expression but without the dot operator. This prevents recursion.
Then, when the non-dot expression has been parsed, check if a dot and an identifier follows. If this is not the case, we are done. If this is the case, wrap the current node up in a dot operation node. Then, keep track of the entire string text that has been parsed for this operation so far. Then revert everything back to before the operation was being parsed, and now re-parse a "custom expression", where the first directly-nested expression would really be trying to match the exact string that was parsed before rather than a real expression. Repeat until there are no more dot-identifier pairs (this should happen automatically by the new "custom expression").
This is messy, complicated and possibly slow, and I'm not entirely sure if it'll work but I'll try it out. I'd appreciate alternative solutions.
I am writing a lexer/parser for a language that allows abbreviations (and globs) for its keywords. And, I am trying to determine the best way to do it.
And one thought that occurs to me, is to insert a phase between the lexer and the parser, where the lexer recognizes the general class, e.g. is this a "command name" or is it an "option" and then passes those general tokens to a second phase which does further analysis and recognizes which command name it is and passes that on as the token type to the parser.
It will make the parser simple. I will only have to deal with well formed command names. Every token will be clear what it means.
It will keep the lexer simple. It will only have to divide things into classes. This is a simple name. This is a glob. This is an option name (starts with a dash).
The phase is the middle will also be relatively simple. The simple name (and option forms) will only have to deal with strings. The glob form can use standard glob techniques to match the glob against the legal candidates, which are in the tables for the simple names and options.
The question is how to insert that phase into ANTLR, so that I call the lexer and it creates tokens and the intermediate phase massages them and then the parser gets the tokens the intermediate phase has categorized.
Is there a known solution for this?
Something like:
lexer grammar simple
letter: [A-Z][a-z];
digit: [0-9];
glob-char: [*?];
name: letter (letter | digit)*;
option: '-'name;
glob: (glob-char|letter)(glob-char|letter|digit)*;
glob-option: '-'glob;
filter grammar name;
end: 'e' | 'end';
generate: 'ge' | 'generate';
goto: 'go' | 'goto';
help: 'h' | 'help';
if: 'i' | 'if';
then: 't' | 'then';
parser grammar simple;
The user (programmer writing the language I am parsing) need to be to write
g*te and have if match generate.
The phase between the lexer and the parser when it sees a glob needs to look at the glob (and the list of keywords) and see if only one of them matches the glob and if so, return that keyword. The stuff I listed in the "filter grammar" is the stuff that builds the list of keywords globs can match. I have found code on the web that matches globs to a list of names. That part isn't hard.
And, I've since found in the ANTLR doc how to run arbitrary code on matching a token and how to change the resulting tokens type. (See my answer.)
It looks like you can use lexerCustomActions to achieve the desired effect. Something like the following.
in your lexer:
GLOB: [-A-Za-z0-9_.]* '*' [-A-Za-z0-9_.*]* { setType(lexGlob(getText())); }
in your Java (or whatever language you are using code):
void int lexGlob(String origText()) {
return xyzzy; // some code that computes the right kind of token type
}
I'm trying to understand how ANTLR grammars work and I've come across a situation where it behaves unexpectedly and I can't explain why or figure out how to fix it.
Here's the example:
root : title '\n' fields EOF;
title : STR;
fields : field_1 field_2;
field_1 : 'a' | 'b' | 'c';
field_2 : 'd' | 'e' | 'f';
STR : [a-z]+;
There are two parts:
A title that is a lowercase string with no special characters
A two character string representing a set of possible configurations
When I go to test the grammar, here's what happens: first I write the title and, on a new line, give the character for the first field. So far so good. The parse tree looks as I would expect up to this point.
When I add the next field is when the problem comes up. ANTLR decides to reinterpret the line as an instance of STR instead of a concatenation of the fields that I was expecting.
I do not understand why ANTLR tries to force an unrelated terminal expression when it wasn't specified as an option by the grammar. Shouldn't it know to only look for characters matching the field rules since it is descended from the fields node in the parse tree? What's going on here and how do I write my ANTLR grammars so they don't have this problem?
I've read that ANTLR tries to match the format greedily from the top of the grammar to the bottom, but this doesn't explain why this is happening because the STR terminal is the very last line in the file. If ANTLR gives special precedence to matching terminals, how do I format the grammar so that it interprets it properly? As far as I understand, regexes do not work for non-terminals so it seems that have to define it how it is now.
A note of clarification: this is just an example of a possible grammar that I'm trying to make work with the text format as is, so I'm not looking for answers like adding a space between the fields or changing the title to be uppercase.
What I didn't understand before is that there are two steps in generating a parser:
Scanning the input for a list of tokens using the lexer rules (uppercase statements) and then...
Constructing a parse tree using the parser rules (lowercase statements) and generated tokens
My problem was that there was no way for ANTLR to know I wanted that specific string to be interpreted differently when it was generating the tokens. To fix this problem, I wrote a new lexer rule for the fields string so that it would be identifiable as a token. The key was making the FIELDS rule appear before the STR rule because ANTLR checks them in the order they appear.
root : title FIELDS EOF;
title : STR;
FIELDS : [a-c] [d-f];
STR : [a-z]+;
Note: I had to bite the bullet and read the ANTLR Mega Tutorial to figure this out.
I am writing a small program which needs to preprocess some data files that are inputs to another program. Because of this I can't change the format of the input files and I have run into a problem.
I am working in a language that doesn't have libraries for this sort of thing and I wouldn't mind the exercise so I am planning on implementing the lexer and parser by hand. I would like to implement a Lexer based roughly on this which is a fairly simple design.
The input file I need to interpret has a section which contains chemical reactions. The different chemical species on each side of the reaction are separated by '+' signs, but the names of the species can also have + characters in them (symbolizing electric charge). For example:
N2+O2=>NO+NO
N2++O2-=>NO+NO
N2+ + O2 => NO + NO
are all valid and the tokens output by the lexer should be
'N2' '+' 'O2' '=>' 'NO' '+' 'NO'
'N2+' '+' 'O2-' '=>' 'NO' '+' 'NO'
'N2+' '+' 'O2-' '=>' 'NO' '+' 'NO'
(note that the last two are identical). I would like to avoid look ahead in the lexer for simplicity. The problem is that the lexer would start reading the any of the above inputs, but when it got to the 3rd character (the first '+'), it wouldn't have any way to know whether it was a part of the species name or if it was a separator between reactants.
To fix this I thought I would just split it off so the second and third examples above would output:
'N2' '+' '+' 'O2-' '=>' 'NO' '+' 'NO'
The parser then would simply use the context, realize that two '+' tokens in a row means the first is part of the previous species name, and would correctly handle all three of the above cases. The problem with this is that now imagine I try to lex/parse
N2 + + O2- => NO + NO
(note the space between 'N2' and the first '+'). This is invalid syntax, however the lexer I just described would output exactly the same token outputs as the second and third examples and my parser wouldn't be able to catch the error.
So possible solutions as I see it:
implement a lexer with atleast one character look ahead
include tokens for whitespace
include leading white space in the '+' token
create a "combined" token that includes both the species name and any trailing '+' without white space between, then letting the parser sort out whether the '+' is actually part of the name or not.
Since I am very new to this kind of programming I am hoping someone can comment on my proposed solutions (or suggest another). My main reservation about the first solution is I simply do not know how much more complicated implementing a lexer with look ahead is.
You don't mention your implementation language, but with an input syntax as relatively simple as the one you outline, I don't think having logic along the lines of the following pseudo-code would be unreasonable.
string GetToken()
{
string token = GetAlphaNumeric(); // assumed to ignore (eat) white-space
var ch = GetChar(); // assumed to ignore (eat) white-space
if (ch == '+')
{
var ch2 = GetChar();
if (ch2 == '+')
token += '+';
else
PutChar(ch2);
}
PutChar(ch);
return token;
}
Im trying to model the EBNF expression
("declare" "namespace" ";")* ("declare" "variable" ";")*
I have built up the yacc (Im using MPPG) grammar, which seems to represent this, but it fails to match my test expression.
The test case i'm trying to match is
declare variable;
The Token stream from the lexer is
KW_Declare
KW_Variable
Separator
The grammar parse says there is a "Shift/Reduce conflict, state 6 on KW_Declare". I have attempted to solve this with "%left PrologHeaderList PrologBodyList", but neither solution works.
Program : Prolog;
Prolog : PrologHeaderList PrologBodyList;
PrologHeaderList : /*EMPTY*/
| PrologHeaderList PrologHeader;
PrologHeader : KW_Declare KW_Namespace Separator;
PrologBodyList : /*EMPTY*/
| PrologBodyList PrologBody;
PrologBody : KW_Declare KW_Variable Separator;
KW_Declare KW_Namespace KW_Variable Separator are all tokens with values "declare", "naemsapce", "variable", ";".
It's been a long time since I've used anything yacc-like, but here are a couple of suggestions that may or may not help.
It seems that you need a 2-token lookahead in this situation. The parser gets to the last PrologHeader, and it has to decide whether the next construct is a PrologHeader or a PrologBody, and it can't tell that from the KW_Declare. If there's a directive to increase lookahead in this situation, it will probably solve the problem.
You could also introduce context into your actions: rather than define PrologHeaderList and PrologBodyList, define PrologRuleList and have the actions throw an error if a header appears after a body. Ugly, but sometimes you have to do it: what appears simple in a grammar may not be simple in the generated parser.
A hackish approach might be to combine the tokens: rather than KW_Declare and KW_Variable, have your lexer recognize the space and use KW_Declare_Variable. Since both are keywords, you're not going to run into namespace collision problems.
The grammar at the top is regular so IIRC you can plot it out as a DFA (or a NDA and convert it to a DFA) and then convert the DFA to a grammar. It's bean a while so I'll leave the work as an exercise for the reader.